Update Substrate to 1.13.0

[nazar-pc]

This updates Substrate from 1.10.1 to 1.13.0.

PR points to tips of https://github.com/subspace/polkadot-sdk/tree/subspace-v6 and https://github.com/subspace/frontier/tree/subspace-v8

Te biggest change upstream was probably related to networking refactoring (they have introduced litep2p library, which is libp2p-compatible reimplementation, as an option).

I have dropped a custom limit for block download response size in Substrate. The reason is that we mostly rely on DSN sync these days and that limit was more important when syncing using Substrate sync during early history with heavy blocks when users had a very slow Internet route to other peers.

Other patches were rebased without major conflicts. I have created two upstream PRs (paritytech/polkadot-sdk#4842 and paritytech/polkadot-sdk#4844) that do not introduce concurrent block verification, but still reduce our diff downstream and shouldn't be very controversial for upstream.

Interesting PRs on Substrate side:

• Renames frame crate to polkadot-sdk-frame paritytech/polkadot-sdk#3813
• Migrate fee payment from Currency to fungible paritytech/polkadot-sdk#2292
• GenesisConfig presets for runtime paritytech/polkadot-sdk#2714 (this might be useful for test network potentially)
• Adds ability to trigger tasks via unsigned transactions paritytech/polkadot-sdk#4075
• rpc: add option to whitelist ips in rate limiting paritytech/polkadot-sdk#3701
• Replace Multiaddr & related types with substrate-specific types paritytech/polkadot-sdk#4198
• Adds ability to specify chain type in chain-spec-builder paritytech/polkadot-sdk#4542
• Add generate and verify logic for AncestryProof paritytech/polkadot-sdk#4430 (@vedhavyas this introduces new proof type in addition to leaf proofs we used before, might be potentially useful)
• [pallet-balances] burn_allow_death extrinsic paritytech/polkadot-sdk#3964

The only potentially interesting PRs on frontier side are these:

• refactor: storage override polkadot-evm/frontier#1425

Code contributor checklist:

• I have read, understood and followed contributing guide

[nazar-pc]

litep2p brings OpenSSL as dependency (paritytech/polkadot-sdk#4856) 😞
Will try to patch polkadot-sdk to get rid of it somehow.

[shamil-gadelshin]

Sync test works.

[nazar-pc]

Created litep2p fork to get rid of OpenSSL: https://github.com/subspace/litep2p/tree/subspace-v1
Reported/commented in various places upstream, so will hopefully be possible to switch back eventually.

[vedhavyas]

Skimmed through MMR leaf proof and Ancestry proof changes.
Since the MMR proofs on Polkadot are limited to just blocks that are not pruned on-chain, They had to introduce the ancestry proof since verifier may not have access to those since it could be verified on Ethereum for example.

We already patched on our end the on-chain block hash limitation and using ancestry proof may not be worth it since it introduces extra storage proofs. On the other hand, we may benefit from not storing the MMR roots since we can use the Ancestry proof to prove any parent blocks.

Will take a closer look again and will initiate some dicussion if we can benefit from the change @nazar-pc. Thanks for pointing it out

[nazar-pc]

Hm... domain tests are failing with too large allocation 🤔
I don't recall changes to WASM heap size. @vedhavyas or @NingLin-P would you mind taking a closer look, please?

[vedhavyas]

@nazar-pc looks like this upgrade increases the genesis payload till 37 Mib while the limit is 32 Mib
A quick solution would be run the tests with release profile so that WASM size is reduced. I tested this localy

I'm still looking if we can build wasm only in release while tests run in Debug profile. If so that would be an ideal route
What do you think ?

[nazar-pc]

IIRC WASM runtime is already build in release even for debug builds 🤔

[vedhavyas]

Apparently not, from the code base, I see we have to explicitly set WASM_BUILD_TYPE_ENV=release to compile to release else it will take root project build type which will be debug.

I'm testing this locally. Will let you know if it works

[nazar-pc]

That is not what I see 🤔 Even for debug builds WASM is built as release:

target/debug/wbuild/subspace-runtime/target/release
target/debug/wbuild/subspace-runtime/target/wasm32-unknown-unknown/release

[vedhavyas]

@nazar-pc Yes, WASM blob is being built with release profile but it skips the compression of compact WASM because tests are run in debug profile. So final WASM_BINARY will include 2.7 Mib wasm instead of 645Kb wasm.

[vedhavyas]

I have the change here - #2872
If CI passes, we can merge it into this @nazar-pc